GLM-5.3-CYBERSECURITY-FP8
> [!IMPORTANT] > Runtime notes — field-tested on 8× DGX Spark GB10 by @0xMagnus (discussion): > > - reasoningeffort only honors "low" and "high". Every other value — off, medium, max, unset, or an unquoted YAML off: (parses as boolean false) — falls through to max. There is no way to disable reasoning on this checkpoint; pass "low" for minimum. > - On FP8, prefer low for agent / tool-loop use. At high/max the model c...
Params
753.3 B
Context
1,048,576
Downloads 30d
130 K
Likes
883
Download history
tracking started — chart appears after 7 days of snapshots (2 recorded)130 K downloads in the last 30 days
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| model.safetensors | f8_e4m3 | 755.6 GB | 944.7 GB | ❌ Won’t fit |
| model.safetensors (bf16, full) | bf16 + 1024K ctx | 755.6 GB | 15,295.6 GB | ❌ Won’t fit |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ curl -s https://aimodelscomparison.com/api/v1/models/glm-5-3-cybersecurity-fp8
{
"hf_id": "dealignai/GLM-5.3-CYBERSECURITY-FP8",
"params_b": 753.33,
"context_length": 1048576,
"license": { "id": "mit", "commercial": "yes" },
"downloads_30d": 130443,
"vram_estimates": [
{ "quant": "f8_e4m3", "gb": 944.7 }
],
"updated_at": "2026-11-09T01:00:18Z"
}
Specifications
- Architecture
- GlmMoeDsaForCausalLM
- Parameters
- 753.3 B
- Tensor type
- F8_E4M3
- Context length
- 1,048,576
- Vocabulary
- 154,880
- Layers / heads
- 78 / 64
- Licence
- mit
- First seen on the Hub
- 2026-08-30
- Base model
- GLM-5.3
- Training datasets
- undisclosed
- Added to our catalog
- 2026-11-09
Compare with any text-generation model